Forty-five percent of women journalists and media workers now self-censor on social media to avoid harassment, a 50 percent jump from 2020, according to a UN Women report released April 30 ahead of World Press Freedom Day on May 3. The study, “Tipping point: Online violence impacts, manifestations and redress in the AI age,” surveyed 641 women human rights defenders, activists and journalists across 119 countries in 2025.

“AI is making abuse easier and more damaging, and this is fueling the erosion of hard-won rights in a context marked by democratic backsliding and networked misogyny,” said Kalliopi Mingerou, who leads UN Women’s team working to end violence against women, in the agency’s announcement.

The findings published by UN News document how generative AI has reshaped the threat landscape. Six percent of respondents reported being targeted with deepfakes, while 12 percent said personal images, including intimate or sexual content, had been shared without their consent. One in three women surveyed said they had received unsolicited sexual advances online.

The professional fallout is substantial. Nearly 22 percent of women journalists said they self-censor in their actual reporting work, retreating from investigative beats under pressure from coordinated harassment campaigns.

“When right-wing groups online brand me a ‘traitor,’ and thousands of WhatsApp forwards spread these false allegations, simply living in my own country becomes terrifying,” an environmental journalist from India told researchers. “We have begun to self-censor, withdrawing from investigative reporting.”

The mental health toll is equally stark. Roughly 24.7 percent of women journalists and media workers reported being diagnosed with or treated for anxiety or depression linked to online abuse, and almost 13 percent have been diagnosed with PTSD, the report found.

One respondent, a journalist and community organizer, told researchers she resigned from her job in December 2023 to focus on her mental health and is now “subsisting on rice porridge, a direct consequence of being forced into silence and out of work.”

There are signs that women in media are pushing back through formal channels. Reports of online violence to police have doubled since 2020, rising from 11 percent to 22 percent. Legal action against perpetrators, tech platforms or employers climbed from 8 percent in 2020 to 14 percent in 2025.

Legal protections, however, remain thin. Fewer than 40 percent of countries have laws shielding women from cyber harassment or cyberstalking, according to World Bank data cited in the report.

For newsrooms, the findings carry direct operational consequences. When nearly half of women journalists are pulling back from social media and one in five is softening their reporting, editors face a measurable loss of source material, beat coverage and audience reach, particularly on investigative and politically sensitive stories. Publishers that have offloaded digital safety to individual reporters now have quantitative evidence that the strategy is producing burnout, resignations and clinical diagnoses. Concrete steps, such as legal support funds, deepfake monitoring tools, takedown protocols with platforms and in-house mental health resources, are no longer optional perks but retention infrastructure.

The UN Women study is the second in a series. The next installment will examine perpetrator behavior and the role of “Big Tech” companies in amplifying or curbing abuse, Mingerou said. With World Press Freedom Day arriving Monday, publishers face a sharpening question: whether the response to AI-enabled harassment will be built into newsroom workflows or left to the women absorbing the damage.

The post AI is supercharging online abuse of women journalists, UN study finds appeared first on The Media Copilot.

“The Epstein Files,” an AI-generated podcast launched in February 2026 by data entrepreneur Adam Levy, has logged more than 2 million downloads by processing over 3 million documents tied to Jeffrey Epstein into a daily, self-updating show hosted by two synthetic voices. The series bills itself as “the first AI native” investigative documentary and presents its output as a “forensic audit,” according to an analysis by Kathryn McDonald, principal academic in audio production at Bournemouth University, published in The Conversation on May 6, 2026.

The pipeline ingests, cross-references and scripts material with no identifiable human speakers behind the hosts, McDonald wrote. Levy has said his goal is to “strip the emotion” from the story, while the show’s hosts claim it combines AI processing with “human analysis” to review the records rather than speculate.

McDonald argues that distinction is hard to verify because the selection, interpretation and emphasis driving the narrative remain largely invisible. The conversational format borrows the cadence of shows like “This American Life,” “Serial” and “S-Town,” complete with jokes, cross-talk, hesitations and filler words. What it lacks, she notes, are interviews, location recordings and meaningful sonic cues.

“Coherence is not the same as sense making, and pattern recognition is not interpretation,” McDonald wrote. Editorial decisions, she argues, do not vanish under automation. They get relocated into training data, system design and weighting mechanisms, then surface as outputs that read as neutral.

The voices themselves do editorial work. McDonald notes the hosts are modeled on familiar broadcast styles tied to authority in Western media, reproducing assumptions about professionalism and trust while remaining detached from any identifiable speaker. The conversational structure suggests multiple perspectives, the tone implies neutrality and the pacing suggests deliberation, none of which guarantees the underlying material has been critically evaluated.

The subject matter sharpens the stakes. The Epstein documents center on human harm and exploitation, McDonald wrote, and stories of that nature demand sensitivity, restraint and a clear chain of accountability. The show offers no visible editorial voice and no apparent right of reply for listeners or subjects, she added.

For newsrooms, the case study cuts in two directions. AI-generated audio is cheap, fast and increasingly hard to distinguish from human-produced work, which means publishers competing on investigative audio now face rivals who can ship daily episodes against document troves no traditional team could process at the same speed. It also hands editors a concrete argument for transparency standards: disclosing who selected sources, who weighed credibility and who is accountable for errors, particularly when synthetic hosts mimic the trusted cadence of established narrative podcasts.

Audio teams should consider publishing methodology notes alongside episodes, naming human editors on AI-assisted productions, and labeling synthetic voices clearly in feeds and show art. Until industry bodies and platforms, including Apple Podcasts and Spotify, set uniform disclosure rules for AI-generated shows, individual publishers will need to decide their own policies for labeling AI-generated content.

McDonald closes with a detail worth sitting with. Listen closely to “The Epstein Files,” she writes, and you will notice that no one ever takes a breath. The next competitive edge in podcasting may not be speed or scale. It may be the audible proof that a human is still in the room.

The post AI-generated ‘Epstein Files’ podcast hits 2 million downloads, raising alarms over invisible editorial judgment appeared first on The Media Copilot.

The tools used to monitor journalists — once confined to intelligence agencies — are now commercially available, widely deployed, and capable of accessing a phone without the target ever clicking a link. On World Press Freedom Day, May 3, the International Federation of Journalists put that reality at the center of its annual assessment of global press conditions, publishing findings that describe not a gradual erosion of media freedom but an accelerating one.

The IFJ, which represents more than 600,000 media professionals across 148 countries, called the global state of press freedom “deplorable.” UNESCO’s latest World Trends Report on Freedom of Expression and Media Development adds the statistical frame: press freedom has fallen 10% since 2012, a decline the IFJ said is comparable to some of the most unstable periods of the 20th century.

128 deaths, and counting

The human cost in 2025 was 128 journalists killed. The IFJ said additional deaths have already been recorded in 2026. Reporters working in conflict zones face the sharpest risks — in Ukraine, Palestine, Lebanon, and Sudan, journalists have been arrested, displaced, or killed while carrying out their work. Individuals identified as press are increasingly becoming deliberate targets rather than incidental casualties.

IFJ General Secretary Anthony Bellanger described each attack as an act with consequences beyond the individual.

“Every attack on a media professional is an attack aimed at silencing a story intended to inform citizens,” Bellanger said, adding that restrictions on journalism ultimately prevent the public from making informed decisions.

Spyware without borders

In a study published April 28 — “Global Surveillance of Journalists: A Technical Mapping of Tools, Tactics and Threats” — the IFJ documented what it describes as a convergence of state intelligence capabilities, private-sector tools, and weak regulatory frameworks.

The report, which draws on cybersecurity expert interviews and technical investigations conducted between 2021 and 2025, identifies commercial spyware systems including Pegasus, Predator, and Graphite as now widely available beyond their original government-intelligence markets. All three are capable of “zero-click” intrusions — accessing a target’s device with no interaction required from the user.

The IFJ found these technologies are frequently deployed with limited oversight, leaving journalists monitored without accountability and with few legal avenues for redress.

AI as force multiplier

The IFJ study also raises concerns about artificial intelligence extending the reach of existing surveillance infrastructure. Data gathered through digital monitoring — communications, location history, online activity — can be fed into AI systems that analyze it at scale. In conflict environments, the report said, such systems can combine telecommunications data with drone feeds, enabling the identification and tracking of journalists in the field.

Beyond targeted surveillance, the IFJ warned of AI-driven disinformation, identity theft, and automated content systems that bypass editorial standards entirely.

Lead study author Samar Al Halal said the effects compound in ways that damage journalism even when no direct harm occurs.

“When journalists are watched, sources disappear, investigations stop, and self-censorship becomes normal,” Al Halal said. “The public doesn’t just lose information, it loses the ability to hold power accountable.”

What the IFJ is demanding

The organization is calling on governments to enact laws protecting press freedom and regulating surveillance technologies, restrict the export and use of commercial spyware, and strengthen legal safeguards for journalists’ sources. The surveillance report specifically recommends increased investment in digital security training and stronger protections for encryption and anonymity.

The broader context makes those demands urgent. A 10% global decline in press freedom over 13 years, 128 journalists dead in a single year, and surveillance tools that require no mistake from their targets — the infrastructure for silencing reporters has rarely been more capable or more available.

The post Spyware and AI surveillance targeting journalist on the rise, IFJ warns appeared first on The Media Copilot.

OpenAI today released GPT-5.5, what it says is its “smartest and most intuitive to use model yet, and the next step toward a new way of getting work done on a computer.”

The company says the model understands what users are trying to do faster, can carry more of the workload itself, and excels at writing and debugging code, researching online, analyzing data, creating documents, operating software, and moving across tools until a task is finished.

The company published performance numbers from Terminal-Bench 2.0, which tests complex command-line workflows requiring planning, iteration, and tool coordination.

OpenAI said GPT-5.5 outperformed its predecessor on every major coding and agent benchmark the company tested, while using fewer tokens and running at the same speed as the older model. On one third-party coding index, it matched leading rivals at about half the cost.

Keeping a larger model that fast required rebuilding inference as a single system rather than a patchwork of tweaks, the company said. GPT-5.5 was designed, trained and served on NVIDIA’s latest hardware, and OpenAI credited its own Codex tool and GPT-5.5 itself with helping hit the efficiency targets.

Early testers told the company the model seems to grasp how a codebase fits together — why something is failing, where the fix belongs and what else the change will touch.

Dan Shipper, Founder and CEO of Every, called GPT-5.5 “the first coding model I’ve used that has serious conceptual clarity.” After launching an app, he spent days debugging a post-launch issue before bringing in one of his best engineers to rewrite part of the system. To test GPT-5.5, he effectively rewound the clock: could the model look at the broken state and produce the same kind of rewrite the engineer eventually decided on? GPT-5.4 could not. GPT-5.5 could.

Pietro Schirano, CEO of MagicPath, saw a similar step change when GPT-5.5 merged a branch with hundreds of frontend and refactor changes into a main branch that had also changed substantially — resolving the work in one shot in about 20 minutes.

One engineer at NVIDIA with early access went as far as to say: “Losing access to GPT-5.5 feels like I’ve had a limb amputated.”

OpenAI is already running the model internally at scale. More than 85% of the company uses Codex every week across functions including software engineering, finance, communications, marketing, data science, and product management. The finance team used GPT-5.5 in Codex to review 24,771 K-1 tax forms totaling 71,637 pages, accelerating the task by two weeks compared to the prior year.

The model also shows gains on scientific and technical research workflows. On GeneBench, a new eval focusing on multi-stage scientific data analysis in genetics and quantitative biology, GPT-5.5 outperforms GPT-5.4 on problems that often correspond to multi-day projects for scientific experts. On BixBench, a benchmark designed around real-world bioinformatics and data analysis, it achieved leading performance among models with published scores.

In a notable example, an internal version of GPT-5.5 with a custom harness helped discover a new proof about Ramsey numbers — one of the central objects in combinatorics — later verified in the Lean proof assistant. The result is a concrete example of GPT-5.5 contributing not just code or explanation, but a novel mathematical argument in a core research area.

OpenAI says GPT-5.5 was released with its strongest safeguards to date, including tighter controls around cybersecurity workflows and repeated misuse patterns. The model was evaluated across the company’s full safety and preparedness frameworks, with input from nearly 200 trusted early-access partners before launch.

GPT-5.5 is available today in ChatGPT and Codex for Plus, Pro, Business, and Enterprise users. GPT-5.5 Pro is rolling out to Pro, Business, and Enterprise tiers. API access, which requires different safeguards, is coming “very soon,” OpenAI said.

The post GPT-5.5 Is ‘Our Smartest Model Yet,’ Says Company With History of Saying That appeared first on The Media Copilot.

The Alliance for Audited Media announced Wednesday it is making its Ethical AI Certification available to all publishers as part of AAM membership. The certification provides a structured framework for developing, implementing, and demonstrating responsible AI governance.

Publishers are navigating rapid AI adoption, emerging regulatory proposals, and rising audience expectations for disclosure and oversight. Recent research from the Local Media Association and Trusting News found that nearly 99 percent of news audiences expect human involvement when AI is used. The finding underscores a challenge for publishers integrating AI tools: audiences notice and adjust their trust based on how disclosure is handled.

AAM developed the certification with publisher input. It evaluates companies across eight key areas including transparency, governance, bias and fairness, and privacy. The program was recently updated to incorporate elements of the IAB’s AI Transparency and Disclosure Framework to strengthen industry alignment, according to the announcement.

Richard Murphy, AAM’s CEO, president and managing director, said trust remains a critical factor in how AI is implemented and used. “By expanding access to our certification, we’re helping publishers demonstrate and communicate responsible AI use to their subscribers, advertisers and partners,” he said.

During the certification process, companies receive feedback on their AI governance policies and oversight mechanisms. Publishers who complete the program can display AAM’s Ethical AI Certification seal on their websites and media kits, and receive a listing on AAM’s Assurance List.

Publishers can find more information at auditedmedia.com.

The post Alliance for Audited Media opens ethical AI certification to publishers appeared first on The Media Copilot.

Most brands have no idea whether ChatGPT recommends them, ignores them, or gets them wrong. Gist — a subsidiary of ProRata, founded by Bill Gross, the entrepreneur credited with inventing the pay-per-click advertising model — launched Gist GEO on April 21 to close that gap.

The platform tracks and improves how brands appear in AI-generated answers across ChatGPT, Google AI Overviews, Gemini, and Perplexity. It is available now at gist.ai/geo.

The problem GEO is built to solve

Traditional SEO tells a brand where it ranks in search results. It says nothing about whether a brand gets cited, recommended, or accurately described when a buyer asks an AI assistant a question. That gap — between SEO visibility and AI visibility — is the market Gist GEO is entering.

The platform measures brand performance across six metrics: AI Mention Share, Share of Citations, Share of Citations Found, Share of Earned Media, Share of Recommendations, and Prominence. Those feed into four Brand Health dimensions — Awareness, Sentiment, Authority, and Recall — updated weekly with directional deltas and trend lines.

Features include an AI Visibility Dashboard, Brand Opportunity Reports, Competitive Benchmarking, Content Gap Analysis, and Strategic Recommendations. The Growth tier ($349/month) covers up to three brand profiles across four AI engines (ChatGPT, Perplexity, Gemini, and Google AI Overviews), with 300 basic tracking queries and 30 deep analysis queries per month. Basic access runs $99/month for a single brand profile, ChatGPT only, with 60 basic tracking queries and 5 deep analysis queries. Enterprise offers unlimited profiles, all AI engines, and custom query volumes at custom pricing. All new accounts start with a 14-day free trial on the Growth tier (credit card required).

Funding and backing

ProRata has raised more than $75 million since its 2024 launch, including a $40 million Series B. The company is partnering with top-tier publishers and positioning Gist GEO as part of a broader suite that includes publisher monetization tools. CMO Peter Sloterdyk is leading the go-to-market push; Gist discussed the launch at the GROW LA roundtable on April 22.

Gross, whose pay-per-click model reshaped online advertising in the late 1990s, is framing Gist GEO as a similarly structural intervention — this time for a world where AI answers are displacing the search results page entirely.

Where GEO fits in the emerging taxonomy

Gist distinguishes GEO from the related but distinct discipline of AEO (Answer Engine Optimization). In its framing: SEO determines where you rank in traditional search; AEO shapes how you appear in AI-generated direct answers; GEO ensures your brand is cited accurately and favorably within the AI narrative. The distinction matters because AI systems don’t just return results — they construct answers, and brands that aren’t structured for AI comprehension can be misrepresented or omitted entirely.

The platform’s “zero-click” framing captures the underlying pressure. AI answers increasingly appear above ranked search results, meaning users never visit a brand’s site at all. A brand that doesn’t show up in that layer may as well not exist for that query.

The post Gist GEO launches to fix AI’s brand blind spot, backed by $75M and the inventor of pay-per-click appeared first on The Media Copilot.

OpenAI has launched ChatGPT Images 2.0, an upgraded image generation model that it says can now reason about visual tasks, render text across multiple languages with high accuracy, and produce up to eight sequential images in a single request.

The update, announced Monday, is available across ChatGPT, Codex, and the API. Advanced features including reasoning-driven outputs are limited to Plus, Pro, and Business subscribers.

The most significant capability for media workflows is what OpenAI calls "text mode" — a distinct generation path optimized for images where written language is a primary design element, not an afterthought. The model has been rebuilt to accurately render small text, iconography, UI elements, and dense layouts. It also produces outputs at up to 2K resolution through the API.

"Images are a language, not decoration. A good image does what a good sentence does — it selects, arranges, and reveals," OpenAI writes in the announcement.

Multilingual support is a major part of the release. Earlier image generation models routinely mangled non-English text, particularly in languages with non-Latin scripts. Images 2.0 adds specific support for Japanese, Korean, Chinese, Hindi, and Bengali, making it possible to generate posters, infographics, and narrative formats where language is integrated into the design.

The reasoning integration sets Images 2.0 apart from previous releases. When used with thinking-enabled models, the generator can analyze complex tasks, pull in real-time information, and produce a coherent set of up to eight images with continuous characters and objects across each frame. OpenAI describes it as moving image generation "from rendering to strategic design, from a tool to a visual system."

Aspect ratio support has expanded to as wide as 3:1 and as tall as 1:3, covering banners, slides, mobile screens, and social graphics without post-processing.

For newsrooms and media teams, the update strengthens the case for AI-assisted visual production. The combination of accurate text rendering, multilingual support, and reasoning-driven sequential outputs brings image AI closer to practical use for infographic creation, explainer visuals, and localized content. OpenAI positions the product as a direct rival to Google's image generation offerings; the Tubi integration with ChatGPT earlier this year offered a preview of how OpenAI is pushing ChatGPT into visual and media workflows beyond text.

OpenAI says ongoing limitations include precise physical reasoning and highly detailed structural diagrams, which may still require human review before publication.

The post ChatGPT Images 2.0 adds text mode, multilingual support, and reasoning-driven outputs appeared first on The Media Copilot.

Artificial intelligence is doing something economists rarely see at once: shrinking employment in affected industries while pushing wages higher. New research from the Federal Reserve Bank of Dallas offers a possible explanation — and it has specific implications for anyone early in their career.

Scott Davis, an assistant vice president in the Dallas Fed's Research Department, analyzed employment and wage data across more than 200 occupations since ChatGPT's release in late 2022. The findings, published February 24, suggest AI is simultaneously replacing entry-level workers and making experienced workers more valuable.

The employment picture

Total U.S. employment has grown about 2.5 percent since fall 2022. Employment in AI-exposed sectors has not kept pace. The computer systems design sector — one of the most AI-exposed in the economy — has shed 5 percent of its workforce. Across the top 10 percent of AI-exposed industries broadly, employment is down 1 percent over the same period.

That decline is landing hardest on young workers. Research from Stanford University's Erik Brynjolfsson and colleagues finds the employment drop in AI-exposed sectors is concentrated among workers under 25. Employment totals for older workers have not declined. According to Dallas Fed economist Tyler Atkinson, the issue isn't layoffs — it's that young workers aren't finding jobs in the first place. The entry-level market in AI-exposed fields is getting much harder to break into, a trend that tracks with AI accounting for 25 percent of U.S. layoffs in March according to Challenger, Gray & Christmas.

The wage picture

Here's where it gets unusual. Despite the employment decline, wages in AI-exposed sectors are rising faster than the national average. Nominal average weekly wages across the economy grew 7.5 percent since fall 2022. In computer systems design, they grew 16.7 percent. Across the top 10 percent of AI-exposed industries, wage growth was 8.5 percent.

Davis found no meaningful relationship between AI exposure and wage growth across 205 occupations — until he added one variable: the experience premium.

The codified vs. tacit knowledge divide

Davis draws on a distinction between codified knowledge — the kind you learn from textbooks — and tacit knowledge, the kind you accumulate through years of practice. His hypothesis: AI can replicate codified knowledge but not tacit knowledge. That means AI substitutes for workers whose primary value is book learning, and complements workers whose value comes from hard-won experience.

Using Bureau of Labor Statistics wage data that separates entry-level and experienced worker pay, Davis calculated an experience premium for each of the 205 occupations. He then tested how AI exposure affected wages differently depending on that premium.

The results were clear. For occupations with a low experience premium — jobs where experienced workers don't earn much more than entry-level workers, like fast-food cooks or ticket agents — increased AI exposure was associated with lower wage growth. AI is substituting for everyone in those roles. For occupations with high experience premiums — lawyers, insurance underwriters, credit analysts, marketing specialists — increased AI exposure was associated with higher wage growth. AI is doing the entry-level work while making expert-level judgment more valuable.

What this means for newsrooms and media teams

The implications run directly through white-collar knowledge work, including journalism and media. The traditional career path — take an entry-level job, do the codifiable tasks, slowly build tacit knowledge — is precisely what Davis says firms are finding cost-ineffective to maintain. That dynamic is already visible in the 2026 journalism layoff wave, which has fallen disproportionately on junior and mid-level roles.

For experienced journalists, editors, and media professionals with deep domain knowledge, the data offer some reassurance. Their tacit knowledge — source relationships, news judgment, contextual understanding — is not easily replicated. For new graduates hoping to learn the craft on the job, the environment is harder. The entry-level work AI can do most easily is often the same work that used to teach people the fundamentals.

Davis doesn't suggest this is permanent. "Leaving new employees off the job ladder is not sustainable in the long run," he writes. AI adoption will require rethinking how entry-level employees develop on the job — but that rethinking hasn't happened yet.

The post AI is shrinking entry-level hiring while boosting pay for experienced workers, Dallas Fed finds appeared first on The Media Copilot.

The National Security Agency is using Anthropic's Mythos AI model despite the Pentagon placing the company on a restricted list, according to multiple media reports published Sunday and Monday.

The NSA's adoption of Mythos—Anthropic's next-generation model known for its advanced coding and autonomous task execution capabilities—puts the intelligence agency at odds with Defense Department officials who have flagged Anthropic as a "supply chain risk" and pushed for phasing out its technology across federal systems. The Information was first to report that the NSA has continued using the model even as the Pentagon's restrictions remain in place.

The dispute between Anthropic and the Defense Department stems from the company's refusal to allow unrestricted use of its AI models in sensitive military and surveillance contexts. Anthropic has resisted deployments involving autonomous weapons and mass surveillance capabilities, leading to the breakdown in relations that prompted the DoD's supply chain designation, according to reports.

For national security agencies, the appeal of advanced AI systems like Mythos appears to outweigh the restrictions. Cybersecurity experts believe such models can identify vulnerabilities and enhance defensive operations—priorities that have kept the NSA turning to Anthropic despite the broader federal backlash. The deployment of Mythos has already drawn scrutiny from other governments; earlier this month, UK and US financial regulators held emergency meetings regarding the model's cybersecurity implications.

The situation illustrates a growing challenge for governments worldwide: balancing rapid AI adoption against security, ethical, and regulatory concerns. As AI systems become more capable and integral to defense infrastructure, policymakers face difficult trade-offs between operational effectiveness and risk management.

The NSA's continued use of Anthropic's technology marks a notable fracture in the otherwise coordinated federal response to the company. While the Pentagon has moved to restrict Anthropic's products from wider defense procurement, intelligence community operations appear to have carved out their own path.

The controversy is unlikely to subside. Congressional scrutiny of the NSA's decision is expected, and technology policy advocates say the episode underscores how quickly AI capabilities are outpacing the regulatory frameworks meant to govern them.

The post NSA Using Anthropic’s Mythos Despite Pentagon Blacklist, Reports Say appeared first on The Media Copilot.

Canva is making its biggest bet since it moved design from desktop software into the browser. On Thursday, the company unveiled Canva AI 2.0 at its annual Canva Create event in Los Angeles — a sweeping overhaul that turns the platform into an agentic, conversational system for getting work done.

The centerpiece is a new architecture layer built on what Canva calls its “frontier AI lab” and years of investment in foundation models for design. Instead of generating a single static output, Canva AI 2.0 maintains context throughout a project, helping users brainstorm, refine, and iterate in a continuous conversation.

“Today’s announcement marks the beginning of the next era of creation,” the company said in its release.

The new system goes live as a research preview on April 16, rolling out to the first one million users who find it on the Canva homepage, with broader access to follow.

What’s new

Canva AI 2.0 introduces four core capabilities:

Conversational Design generates fully editable designs from natural language prompts or dictation. Users describe an idea or goal — no blank page, no template hunting — and Canva AI produces a structured, branded layout. The system stays engaged through the process rather than stopping after the first output.

Agentic Orchestration lets users describe a goal and have Canva AI coordinate the full suite of Canva tools to deliver it. The company’s example: ask for “a multi-channel campaign plan to launch our latest summer products,” and the system generates everything — ready to refine or publish.

Object-Based Intelligence enables precise, targeted edits without disturbing the rest of a design. Swap an image, change a headline, adjust a font — only that element changes, and everything remains layered and fully editable.

Living Memory builds a persistent profile of how a user or team works. The system learns preferences, keeps designs on brand automatically, and gets more tailored with use. Users can also seed it with existing designs to create a custom memory library.

Workflows and integrations

Beyond design generation, Canva AI 2.0 introduces several workflow tools aimed at replacing the patchwork of apps most teams currently use:

Connectors link Canva AI to Slack, Gmail, Google Drive, Notion, Zoom, and Google Calendar. The system can pull from those data sources to generate meeting summaries from Zoom transcripts, turn customer emails into sales pitches, or build newsletters from Slack activity.

Scheduling lets users set recurring tasks — generate a week’s worth of social content every Friday, translate it into ten languages, have briefing docs ready at login — and Canva AI runs them automatically in the background.

Web Research brings research directly into designs. Users can run on-demand searches or schedule background research, and Canva AI delivers structured, editable content into the design — no copy-paste required.

Brand Intelligence enforces brand standards automatically across every new design, applying fonts, colors, and style without manual intervention. It can also reapply updated brand guidelines to existing work in a single step.

Canva Code 2.0 now supports HTML importing. Users can bring any HTML file or AI-generated experience into Canva and edit it visually — no code rebuilds needed. The resulting interactive content can include forms that feed into Canva Sheets, or be published to a custom domain with SSO protection.

Sheets AI generates fully structured, data-populated spreadsheets from a description. Budget trackers, project timelines, content calendars, research tables — delivered already formatted.

Why it matters for media teams

The combination of connectors, scheduling, and agentic orchestration makes Canva AI 2.0 a significant tool for editorial and communications teams. The ability to pull from Slack, Gmail, and calendar data and generate campaign materials, briefings, or social content automatically — on a schedule, in the background — is a meaningful reduction in manual production work.

The persistent memory and brand intelligence features address a consistent pain point: keeping output on-brand without manual QA on every piece. For teams managing high-volume content across channels, that’s not a minor efficiency gain.

Canva serves more than 250 million monthly active users across more than 190 countries. The research preview launches April 16.

The launch comes a day after Adobe announced its own agentic creative assistant, Firefly AI Assistant, which similarly orchestrates multi-step workflows across Creative Cloud apps from a conversational interface.

The post Canva launches AI 2.0 with agentic orchestration appeared first on The Media Copilot.