Britain’s competition for artificial-intelligence researchers and other scarce technology specialists is adding pressure to a long-running employment-law debate, as ministers consider whether non-compete clauses should be capped, restricted by salary, or made unenforceable altogether.

The government reopened the issue this year through a working paper examining several reform options after earlier plans for a three-month statutory cap were never implemented. The current options include a universal time limit, different limits according to employer size, a salary threshold below which non-competes would be banned, and an outright prohibition.

Technology companies and investors have a particular interest because employee mobility can determine how quickly expertise moves between incumbents, scaleups, and new businesses. The Competition and Markets Authority has backed tighter restrictions, including a ban below a salary threshold combined with a statutory limit above it.

Industry pressure is also extending beyond conventional non-compete clauses towards lengthy periods of garden leave, under which employees remain on payroll but are prevented from joining another employer. That distinction complicates reform because a restriction can have a similar effect on mobility even when it is structured differently in the contract.

AI turns mobility into a competition issue

The rise of generative AI has intensified competition for engineers, researchers, product specialists, and commercial staff with experience deploying complex systems. When a relatively small pool of workers holds scarce expertise, restrictions on movement can affect not only individual careers but also where new teams form and which businesses can recruit them.

Government evidence suggests non-compete clauses are especially prevalent in higher-paying sectors, including information and communications, although they also appear in lower-paid work. That broad use is one reason ministers are considering whether a single rule would protect mobility effectively or simply create a new default period that employers treat as automatically acceptable.

A statutory cap could reduce the longest restrictions without removing them entirely, while a salary threshold could target protection towards workers considered less able to absorb a period outside their profession. A complete ban would create greater certainty, although it would also remove a tool employers use to protect customer relationships, strategic plans, and confidential information after senior staff leave.

Restrictions can protect investment as well as incumbents

Startups and scaleups do not have a uniform interest in weaker non-competes. Smaller technology companies can benefit when experienced staff become easier to hire from larger competitors, but they can also rely on restrictive covenants when investing heavily in employees with access to code, product roadmaps, customer data, or commercially sensitive research.

Existing UK law already requires a non-compete to be reasonable and connected to a legitimate business interest if it is to be enforceable. The difficulty is that testing reasonableness through litigation can be expensive, leaving workers and smaller companies to make decisions under uncertainty even where a clause might ultimately fail in court.

That is why the reform debate is partly about the practical effect of contracts rather than their formal enforceability. A restriction does not need to survive a full trial to influence whether an employee accepts a competing offer or whether a prospective employer is prepared to risk a dispute.

Garden leave complicates reform

Garden leave presents a harder policy problem because the employee continues to receive salary and benefits while remaining unavailable to competitors. From an employer’s perspective, that can provide time for sensitive information to lose value or customer relationships to transfer, while avoiding some of the financial hardship associated with an unpaid non-compete.

For fast-moving technology markets, however, six or twelve months away from active work can still have a substantial career and competitive effect. Skills, products, and research programmes can move quickly enough that a long paid absence may reduce mobility almost as effectively as a post-employment restriction.

Any reform limited narrowly to contractual non-competes could therefore change drafting behaviour without changing the underlying competition for talent. Employers might lean more heavily on notice periods, garden leave, confidentiality provisions, or other contractual protections if one form of restriction becomes less useful.

The government has not yet chosen its preferred model, and the trade-offs are sharper in technology than a simple employer-versus-worker framing suggests. Reform will have to distinguish legitimate protection of sensitive business interests from restrictions that mainly make it harder for experienced people to take their knowledge and skills into competing companies.