Tag: neural-data

6 articles

Policy & Regulation

UN Secretary-General recommends banning coercive use of neurotechnology in justice systems

In a report to the UN General Assembly dated 27 August 2026 (A/81/351), Secretary-General António Guterres recommends that member states prohibit the compulsory or coercive use of neurotechnology in the administration of justice, including brain-based deception detection, predictive applications, decoding applications that claim to read thoughts, intentions or memories, neuromodulation, and arbitrary or unlawful cognitive surveillance and monitoring. He asks states to recognise neural data as highly sensitive personal data with the highest level of legal protection and clear limits on its collection, processing, sale, storage, retention, reuse and sharing, to require free, prior, informed, explicit and revocable consent for any use of neurotechnology in justice settings, and to urge neurotechnology companies to carry out human rights due diligence. The report, requested by the General Assembly in December 2024, warns, citing outside research, that neural data gathered by consumer devices may be accessed by state authorities or introduced as evidence. Its recommendations do not bind governments.

Oct 11

Market Moves

Sabi raises $50 million for a brain-sensing baseball cap

Khosla Ventures has led a $50 million seed round in Sabi, a California startup building a baseball cap meant to turn imagined words and trained mental commands into text and instructions for AI assistants. Accel, Initialized Capital, former OpenAI chief product officer Kevin Weil, DST Global, Collaborative Fund and Ascend also invested, Sabi announced on 9 October 2026. Forbes put the valuation at $600 million, citing a person familiar with the deal. The cap is designed to read brain signals through hair with sensors built around a custom EEG chip, and to decode them with a model Sabi says was trained on more than 100,000 hours of labelled recordings. In April, Sabi said its first product would be a beanie available by the end of 2026. The cap will now debut as an early prototype at CES in January, and Forbes reported that it will not be available until 2027. Sabi has named no price and published no peer-reviewed results, and its only accuracy figure, up to 77 per cent, comes from a wired laboratory prototype.

Oct 11

Industry News

Precision Neuroscience posts a 70-patient map of the human nervous system

Precision Neuroscience announced on 8 October 2026 a preprint, posted on bioRxiv on 1 October, that describes an atlas of the human central nervous system built from recordings in the first 70 patients fitted with its Layer 7 thin-film electrode array. The array sits on the surface of the brain and holds FDA 510(k) clearance for temporary use of less than 30 days. The atlas spans motor, sensory, language and auditory cortex, and also the brainstem and spinal cord, across more than 100,000 electrode placements. Precision says the total number of patients implanted with its system has now passed 100. Its chief science officer, Benjamin Rapoport, describes a whole-brain interface as a data milestone assembled across many patients, and its chief data and AI officer, Craig Mermel, says the company will soon say more about where the data meets artificial intelligence. The preprint has not been peer reviewed, and 16 of its 61 authors work for Precision.

Oct 9

Industry News

Neuralink cuts some implant users' calibration from 10 minutes a day to 10 minutes a week

Neuralink said on 1 October 2026 that participants in its clinical trial have streamed more than 50,000 hours of unlabelled brain recordings during everyday use of its implant, and that it has used participants' own recordings to pretrain participant-specific models that steady the software turning brain signals into cursor movement. Calibration had averaged 55 minutes a week across participants, typically 10 minutes at the start of each day; some participants now calibrate for 10 minutes a week. Six participants set personal bests on Neuralink's cursor test, and one reached 11.32 bits per second, which the company calls a record for a brain-computer interface. Every live result came from a model trained on a single participant's data, and models pooled across participants have not yet performed better. Neuralink's stated next goals are calibrating once, never calibrating, and a shared 'API for the motor cortex' for developers.

Oct 3

Policy & Regulation

California bans employers from using AI tools to collect workers' neural data

California Governor Gavin Newsom signed AB 1883 on 30 September 2026, the last day he could act on bills this session. From 1 January 2027 the law, by Assemblymember Isaac Bryan (D-Los Angeles), bars employers from using workplace surveillance tools that use artificial intelligence to collect employees' neural data or to recognise, infer or predict their emotional state. It defines neural data as information generated by measuring activity in an employee's central or peripheral nervous system, covers public as well as private employers, and allows penalties of up to $500 per violation, enforced by the Labor Commissioner and public prosecutors. Tools used to ensure safety, and some federally regulated aerospace and defence work, are exempt. In his signing message Newsom said the law lacks key definitions and asked the Legislature to refine it. Three days earlier he vetoed AB 1542, leaving consumers with an opt-out over the sale of the same category of data.

Oct 2