Meta Platforms is making waves in robotics with its latest tactile AI technology, but not without intensifying scrutiny over its approach to open-source AI. Last week, Meta revealed its Digit 360 tactile sensor, developed in partnership with GelSight Inc., alongside advancements in robotic dexterity through the Allegro Hand, created with Wonik Robotics.
These innovations aim to close the gap between human and robotic touch, yet critics argue Meta’s overall AI practices undermine principles of openness and transparency.
Meta’s Tactile Sensing Breakthroughs
Meta’s latest AI advancements bring human-like touch capabilities to machines. The Digit 360 sensor, officially announced with GelSight on October 31, uses sophisticated optical systems to detect pressure, vibrations, heat, and even chemical properties, replicating the sensitivity of human touch. With an AI chip built in, Digit 360 enables robots to react instantaneously, a crucial feature for applications ranging from robotic surgery to virtual reality.
The Allegro Hand, set to launch in 2025, integrates Digit 360 sensors across its joints and fingertips, elevating the precision of robotic grasp and manipulation.
Tactile sensing, the technology behind these innovations, allows robots to interpret and react to physical interactions in real time. As the field evolves, it is increasingly seen as key to enabling robots to perform complex, dexterous tasks in unstructured environments.
Open Source Debate: Meta Under Fire
While Meta’s robotics division advances, its general AI ethics are under scrutiny. On October 28, the Open Source Initiative (OSI) launched the Open Source AI Definition (OSAID), which spells out the requirements for true open-source AI. According to OSAID, an AI model must allow unrestricted use, modification, and distribution, and fully disclose its training data. Meta’s LLaMA models fail to meet these criteria, and OSI has labeled Meta’s open-source claims as misleading.
Stefano Maffulli, OSI’s executive director, has been vocal in his criticism. He says that Meta’s licensing undermines open-source values, creating confusion and limiting user control. Despite widespread adoption of the LLaMA models, Meta’s restrictions, such as bans on commercial use and non-disclosure of training datasets, continue to draw backlash.
Meta argues that its limitations are necessary safeguards, pointing to the models’ potential for misuse. CEO Mark Zuckerberg has repeatedly framed Meta as an AI-first company committed to openness. However, a 2023 study from Radboud University found that many AI models from companies like Meta restrict access to key data, undermining true transparency.
Spot AI: Surveillance Tech Raises Ethical Concerns
Simultaneously, the ethical debate extends to other tech developments, such as Spot AI’s Video AI Agents. The San Francisco-based startup secured $31 million in Series B funding from Qualcomm Ventures to advance its intelligent surveillance systems. These agents, which use edge computing to process data locally, can halt machinery to prevent accidents or alert staff to long queues. The real-time nature of edge computing minimizes latency, but also raises concerns about data privacy and potential misuse.
Spot AI’s technology, while innovative, is compared to dystopian scenarios from 1984 and Minority Report, given the risks of automated surveillance. Critics warn of governments or corporations using these systems to enable mass surveillance, sparking fears of Orwellian control. CEO Rish Gupta has defended the technology’s safety benefits but acknowledged its dual-use potential, saying the were aware of the concerns and are committed to responsible development.
Regulatory Pressure and AI Transparency
The pressure on companies to be transparent and ethical in AI practices is mounting. In late October 2024, the European Commission tightened regulations to demand greater AI accountability, in line with OSI’s new standards. Policymakers stress that AI systems must be understandable and transparent to both regulators and the public, a stance that poses challenges for companies like Meta.
