OpenAI and Meta Push AI Agents Despite 85.5% Trust Gap

In September 2026, OpenAI and Meta launched their respective AI agents, dots and Muse, racing to integrate autonomous software into the lives of hundreds of millions of users. The stakes are not just technical but deeply rooted in public trust. Bloomberg’s October 1 report, “OpenAI and Meta Push Ahead With AI Agents, Testing Public’s Trust,” highlights this critical juncture as two of the world’s leading AI companies ask consumers to hand over significant personal information.

Launch and Scope

OpenAI unveiled dots on September 29, 2026, at its developer conference in San Francisco. The new agents are described as “always-on,” capable of pursuing a user’s goals across multiple apps without waiting for explicit prompts. This framing is significant because while a chatbot answers a question, an always-on agent continues to work after the conversation ends, handling tasks like checking email, rescheduling meetings, and following up on tasks.

CEO Sam Altman emphasized that dots had undergone safety testing and were powered by GPT-6 Astra, OpenAI’s top tier system. The agents require explicit user consent before taking sensitive actions, such as changing passwords. OpenAI has already rolled out dots to ChatGPT Pro, Business Premium, and Enterprise subscribers, with ChatGPT Pro priced at $100 a month. Access to the agents is tied to the highest paying tiers first, as is a familiar OpenAI pattern.

Meta’s Muse, launched in September, is a free agent with a usage limit, starting at $20 a month for paid access. TechCrunch reported that Muse’s download trajectory is impressive, surpassing ChatGPT’s early growth curve. However, the skepticism around Muse’s launch is evident, as users must trust Meta with extensive personal information.

Trust and Skepticism

The 4,000 app figure is significant. An agent that can only touch a few tools is a novelty, but an agent that can interact with four thousand tools represents a layer of software sitting atop a user’s entire digital life. OpenAI has already announced texting as an upcoming feature, which would further extend the agent’s reach into one of the most personal communication channels.

Meta’s Muse faces additional scrutiny. Bank of America analysts noted that while feedback around Muse has been positive, privacy and trust remain key considerations for broader adoption. Amazon reportedly blocked Meta’s Muse from certain integrations due to security concerns, highlighting the ongoing trust issues.

Technical and Regulatory Challenges

The transition from chatbots to agents introduces new risks. OpenAI has had to publicly acknowledge agent related incidents, including image leaks and interactions with government agencies. These incidents underscore the broader challenge of agent software failing in ways a chatbot cannot. Reversibility is critical, and OpenAI’s consent requirement for sensitive actions reflects this.

Survey data shows that 85.5% of engineers trust agent output, but this trust is not universal. Consumer trust is a higher bar, and both companies are operating in a regulatory landscape where the Federal Trade Commission is scrutinizing how AI agent products handle consumer protection. The White House AI accord has made outside audits an explicit expectation for frontier AI deployments.

Business and Strategy

Agents represent a significant shift for both companies. OpenAI’s dots extend into enterprise workflows, potentially competing with Microsoft’s established position. Meta’s Muse is a hedge against slower growth in its core advertising business, providing a second product line with potential subscription revenue.

For enterprise customers, the rollout is staged to build usage data and minimize disruption. OpenAI’s decision to gate dots behind higher paying tiers is both a business and safety choice, allowing the company to gather real usage data before expanding. Meta’s free with paid upgrade model maximizes reach quickly but faces greater scrutiny from individual users.

Conclusion

The public trust gap is a critical challenge for both OpenAI and Meta. The launch of agents marks a significant shift in how AI is integrated into daily life, requiring users to grant standing permissions. As the technology advances, the stakes will only continue to grow, demanding robust safety measures and transparent communication to build and maintain public trust.

Source: https://shattered.io/openai-meta-ai-agents-85-percent-trust-gap-2026/