🌈 GLAAD's AI warning
1. Key Themes
Theme 1: AI Bias as a Safety and Trust Crisis — Not Just an Ethics Issue
Biased AI training data poses concrete real-world risks to marginalized groups, and foundational model errors cascade downstream across dependent systems.
"AI models can detect who you are by the questions you ask and what you're doing, and in 60 countries it's still illegal to be LGBTQ+. So if you're criminalized in a country and it's now identifying you as LGBTQ, there is some real risk there."
"She said errors in proprietary models such as ChatGPT, Gemini and Claude can spread across dependent systems. Therefore, those foundational models need to ensure their material is accurate."
Theme 2: Trust as the Primary AI Competitive Moat
As AI assistants proliferate, companies that win user trust will win the market. This frames trust not as a soft value but as the core strategic asset.
"Trust is the biggest currency with AI, and whoever wins this trust war wins the AI war." — Sarah Kate Ellis, GLAAD CEO
"Yahoo is betting on its trust and expertise, and building new experiences to make it more accessible" as more people turn to ChatGPT, Gemini and other AI assistants.
Theme 3: IP Licensing for AI — A New B2B Revenue Category Emerges
Consumer brands are actively building infrastructure to monetize their intellectual property in the AI era, shifting from a posture of passive IP loss to proactive licensed deployment.
"It is built around a creator-first model that gives voice talent and creatives a meaningful seat at the table. It gives brands a trusted way to bring characters into new AI-enabled platforms without losing what makes them authentic." — Hasbro CEO Chris Cocks
"The new effort seeks to change today's default, where AI image and video generators often surface copyrighted characters, but without permission or compensation to the people and companies that create them."
Theme 4: Vertical AI — Domain-Specific Experiences Are the Differentiation Play
Generic AI assistants are losing ground to domain-specific AI products built on proprietary data and expert voices, particularly in high-engagement verticals like sports and finance.
Yahoo's "Ask Kevin O'Connor" feature "lets people ask questions and receive AI-generated responses based on O'Connor's analysis and written voice," pulling from "his NBA Draft Guide, coverage and show along with player and team stats."
Yahoo ranks as "the third-largest U.S. web property, with 250 million monthly users in the U.S., and 75% of its impressions are logged-in users."
Theme 5: SpaceX IPO as a Bellwether for AI-Infrastructure Capital Markets
SpaceX's record-breaking $75B IPO signals that AI infrastructure plays — not just software — are attracting historic capital, with space-based data infrastructure positioned as the next frontier.
"This would be the largest IPO ever, and it's not even close. The current global record is $29.4 billion by Saudi oil giant Aramco in late 2019."
"Its bigger bet is AI and orbital data centers... The company argues the AI sector offers $26.5 trillion in total addressable market (TAM) — a figure that's impossible to verify. For context, the entire U.S. gross domestic product is around $32 trillion."
2. Contrarian Perspectives
Perspective 1: Fixing AI bias is cheap — and companies are just choosing not to do it. The conventional narrative treats AI bias as technically difficult and expensive. Ellis reframes it as a resource allocation choice: a fractional investment from AI companies would be sufficient to meaningfully address the problem.
"If you just think about putting 1% of the money that all these folks are raising to help fix the foundation of this, it would go an enormous way to make this a safer platform, not just for LGBTQ people, but for everyone and every family."
The implication: AI safety failures at the foundational level are not capacity constraints — they are priority decisions.
Perspective 2: SpaceX's real business is AI infrastructure, not space. Most people still think of SpaceX as a launch company. The IPO framing reveals the company is repositioning itself as an AI and orbital data infrastructure play — with a TAM claim that dwarfs its current revenue base.
"SpaceX began as a space launch company and currently makes most of its money from the Starlink connectivity service. Its bigger bet is AI and orbital data centers."
The company claims "$26.5 trillion in total addressable market (TAM) — a figure that's impossible to verify. For context, the entire U.S. gross domestic product is around $32 trillion."
The contrarian read: SpaceX's valuation is not a space bet — it's an AI infrastructure bet, priced accordingly.
Perspective 3: Yahoo may be more competitively positioned in the AI era than commonly assumed. Yahoo is widely considered a legacy internet brand in terminal decline. However, its 250M logged-in monthly U.S. users and domain-specific data assets in sports and finance give it structural advantages that pure-play AI assistants lack.
"75% of its impressions are logged-in users" — a first-party data advantage that most AI competitors cannot match.
CEO Jim Lanzone signaled an IPO is coming: "The plan is not to stay private forever... I don't think it's too far off."
3. Companies Identified
- Description: LGBTQ+ media advocacy organization
- Why mentioned: CEO warned at Axios AI+ Summit that biased AI training data creates real-world dangers for LGBTQ+ communities and previewed an upcoming report on AI bias reduction
- Quote: "What that means, though, is that as these AI models are learning, they're learning with this information and rhetoric coming into them."
Yahoo
- Description: Legacy internet portal, owned by Apollo Global Management since 2021; third-largest U.S. web property
- Why mentioned: Launching two vertical AI products ("Ask Kevin O'Connor" for NBA, "Ask Yahoo Scout" for finance) powered by its Scout AI engine; signaling IPO readiness
- Quote: "The plan is not to stay private forever... I don't think it's too far off." — CEO Jim Lanzone
- Description: Major toy and entertainment company (Mr. Potato Head, Transformers, Clue, G.I. Joe franchises)
- Why mentioned: Launching "Sixth Wall," a new B2B unit to license AI versions of 11 characters, partnering with ElevenLabs for authentic voice replication
- Quote: "It gives brands a trusted way to bring characters into new AI-enabled platforms without losing what makes them authentic." — CEO Chris Cocks
ElevenLabs
- Description: Voice AI specialist company
- Why mentioned: Named as Hasbro's technical partner for recreating authentic character voices for AI licensing
- Quote: "The toymaker is working with real actors and ElevenLabs, which specializes in voice AI, to ensure that the characters have the voice that fans have come to know in movies and TV shows."
SpaceX
- Description: Private aerospace and connectivity company founded by Elon Musk; planning the largest IPO in history
- Why mentioned: Filing to raise $75B at a $1.75 trillion valuation, pivoting its growth narrative to AI and orbital data centers
- Quote: "It would be valued at about $1.75 trillion at that price."
- Description: Parent company of Google and DeepMind
- Why mentioned: Selling $80B in shares to fund AI data centers; also giving UK publishers an opt-out from Google AI Overviews
- Quote: "Alphabet will sell $80 billion in shares to help pay for its AI data centers."
Mattel
- Description: Rival toy company (Barbie, Hot Wheels)
- Why mentioned: Cited as a comparable case — announced OpenAI partnership for AI toys in 2025 but has yet to release any products, providing context for Hasbro's move
- Quote: "Mattel said last year it was working with OpenAI on toys for older kids using its franchises, but has yet to unveil any products."
IBM
- Description: Enterprise technology company
- Why mentioned: CEO Arvind Krishna publicly backed Trump's new AI executive order, endorsing light government guardrails over active regulation
- Quote: "IBM CEO Arvind Krishna backs Trump's new AI executive order, preferring light government guardrails over more active intervention."
Apollo Global Management
- Description: Private equity firm
- Why mentioned: Current owner of Yahoo since 2021; context for Yahoo's private status and pending IPO discussion
- Quote: "Apollo Global Management acquired Yahoo in 2021."
4. People Identified
Sarah Kate Ellis
- Description: CEO of GLAAD
- Why mentioned: Delivered keynote warning at Axios AI+ Summit about AI bias risks to LGBTQ+ communities; previewed an upcoming GLAAD AI bias report
- Quote: "Trust is the biggest currency with AI, and whoever wins this trust war wins the AI war."
Jim Lanzone
- Description: CEO of Yahoo
- Why mentioned: Announced two new vertical AI products at Axios AI+ Summit and signaled Yahoo's IPO is approaching
- Quote: "The plan is not to stay private forever... I don't think it's too far off."
Kevin O'Connor
- Description: Yahoo Sports NBA Senior Analyst; has covered the NBA since 2013
- Why mentioned: The subject of Yahoo's "Ask Kevin O'Connor" AI product — his analysis and written voice are used to power AI-generated responses for NBA Draft queries
- Quote: "O'Connor joined Yahoo Sports as NBA senior analyst in 2024 and has covered the league since 2013."
Chris Cocks
- Description: CEO of Hasbro
- Why mentioned: Announced Sixth Wall licensing unit and articulated the creator-first philosophy behind Hasbro's AI character licensing strategy
- Quote: "It is built around a creator-first model that gives voice talent and creatives a meaningful seat at the table."
Arvind Krishna
- Description: CEO of IBM
- Why mentioned: Publicly endorsed Trump's AI executive order favoring light regulatory touch
- Quote: "IBM CEO Arvind Krishna backs Trump's new AI executive order, preferring light government guardrails over more active intervention."
Elon Musk
- Description: Founder and CEO of SpaceX
- Why mentioned: SpaceX's historic $75B IPO filing centers on his company's pivot to AI and orbital data centers
- Quote: "Musk's company argues the AI sector offers $26.5 trillion in total addressable market (TAM)."
5. Operating Insights
Insight 1: Build vertical AI products around proprietary expert voices and data — not generic models. Yahoo's strategy of anchoring AI answers to a named analyst's (Kevin O'Connor's) existing body of work — rather than a generic AI response — is a replicable playbook for any content or data business. Domain-specific, attributed AI experiences can differentiate against ChatGPT and Gemini by offering trust and credibility that general models cannot.
"Released for the NBA Draft, Yahoo Sports' 'Ask Kevin O'Connor' feature lets people ask questions and receive AI-generated responses based on O'Connor's analysis and written voice... The responses pull from his NBA Draft Guide, coverage and show along with player and team stats."
Insight 2: Foundational model errors compound — companies building on top of AI must audit upstream bias, not just their own outputs. For operators deploying AI in any customer-facing context, the risk is not just internal model behavior but inherited errors from the foundational models they depend on. This is a due-diligence and vendor-selection issue, not merely a technical one.
"She said errors in proprietary models such as ChatGPT, Gemini and Claude can spread across dependent systems. Therefore, those foundational models need to ensure their material is accurate."
Insight 3: Pursue B2B licensing before B2C products when entering AI with IP assets. Hasbro's deliberate choice to go B2B with Sixth Wall — rather than launch consumer AI products — limits execution risk, generates recurring licensing revenue, and avoids the reputational and regulatory exposure of deploying AI directly to consumers (especially children).
"Hasbro is pursuing a business-to-business strategy rather than creating new consumer products."
"Hasbro said it is not aiming its products at those under 13 and is still exploring the best way to bring AI-equipped toys to younger audiences."
6. Overlooked Insights
Insight 1: UK publishers can now opt out of Google AI Overviews — a potentially significant precedent for content creator leverage. Briefly noted in the newsletter's news roundup, this opt-out mechanism — prompted by publisher complaints about traffic drops — signals a potential regulatory and market shift in how AI search products must negotiate with content creators. If adopted more broadly, it could accelerate the push for compensation frameworks between AI companies and publishers.
"Online publishers in the U.K. can opt out of appearing in Google's AI Overviews, after complaints of drops in traffic."
Insight 2: Hasbro is actively engaged in "broader industry discussions around safety standards and voluntary guardrails for AI-enabled play experiences" for children. This single line hints that an industry-wide standard-setting process for AI in children's products may be underway — a nascent but potentially influential regulatory and market-shaping development that received no further elaboration.
"Hasbro said it is not aiming its products at those under 13 and is still exploring the best way to bring AI-equipped toys to younger audiences as well as engaging in 'broader industry discussions around safety standards and voluntary guardrails for AI-enabled play experiences.'"