NewsToolsGuidesExplainedCommunity
AI News

The Path, founded by Tony Robbins and Calm alums, hopes to offer safer AI therapy

The Path says its AI model has scored 95 on the mental health safety AI benchmark, Vera-MH. This compares to a top score of 65 for the consu

2026-05-21 4 min read Marcus J.
The Path, founded by Tony Robbins and Calm alums, hopes to offer safer AI therapy

Most surprising fact: An AI therapy platform, The Path, built by former Tony Robbins and Calm executives, claims a 95 score on a mental health safety benchmark – a score significantly higher than any consumer-facing AI chatbot currently available. This seemingly improbable achievement immediately raises serious questions about the very nature of safety in rapidly developing artificial intelligence, particularly when applied to sensitive areas like mental wellbeing. It’s a claim demanding intense scrutiny, and one that could dramatically shift the landscape of how we approach AI-driven support.

The Path, founded by former Calm CEO Andy Siegel and a team including individuals who previously worked alongside Tony Robbins, is aiming to provide a more secure and reliable experience than existing AI therapy options. Launched this month, the platform utilizes a proprietary AI model trained on a massive dataset of therapeutic conversations and designed to identify and mitigate potentially harmful responses. It’s currently offering a subscription service focused on cognitive behavioral therapy (CBT) and mindfulness techniques, initially targeting stressed professionals. Siegel and The Path’s leadership team are positioning the platform as a significant step toward responsibly integrating AI into mental healthcare, emphasizing a commitment to rigorous testing and ongoing monitoring.

What This Actually Means

Why this matters versus before is profound. Until now, the consumer AI therapy market has been largely unregulated and dominated by chatbots offering generic advice or automated exercises. These bots, like Woebot and Replika, have scored poorly on established safety benchmarks, demonstrating a worrying tendency to escalate distress, provide inappropriate advice, or even exhibit concerning biases. The Path's score of 95 on the Vera-MH benchmark, a standard developed by researchers at the University of Southern California, represents a genuine leap forward – suggesting a vastly improved ability to detect and prevent potentially damaging interactions. However, the benchmark itself is still relatively new, and its limitations are already being debated.

Real-world impact for people and businesses is potentially enormous. For individuals struggling with anxiety, depression, or other mental health challenges, a more reliable and safe AI companion could offer accessible support 24/7. Businesses could also leverage The Path’s technology to provide mental wellbeing resources to employees, potentially reducing burnout and improving productivity. Furthermore, the platform’s success could incentivize further investment in robust AI safety research and development within the broader mental healthcare industry, moving beyond superficial marketing claims.

Bigger picture in the AI race: The Path’s achievement isn’t just about one company; it's a critical test case for the entire field of AI therapy. The industry is currently racing to develop sophisticated AI models capable of simulating human conversation and providing personalized support. However, the ethical and safety considerations are lagging behind. The Path’s focus on a safety benchmark highlights a growing recognition that technical prowess alone isn’t enough—developers need to prioritize responsible design and rigorous evaluation. Competitors like BetterHelp, which has a significant market share, will undoubtedly be pressured to demonstrate comparable safety measures.

Why This Changes Everything

What to watch next: Specifically, AIZyla.com will be closely monitoring The Path's ongoing performance on the Vera-MH benchmark over the next six months. We’ll also be scrutinizing the transparency of their data collection and training methods, demanding to see the full dataset used and the specific algorithms employed. Equally important is observing how The Path handles user feedback and adapts its model in response to identified risks. Finally, we’ll track the development of competing safety benchmarks and the broader industry’s response to The Path's success – will it lead to a universal standard for AI mental health tools, or will it simply become another marketing buzzword?

Stay updated: Follow AIZyla for daily AI news explained clearly for everyone.

Stay ahead of AI -- free

Weekly digest of the best AI news, tools, and guides. No spam.