FOLLOWING THE news of the leaked European Union (EU) regulatory proposal set to tackle issues surrounding high-risk Artificial Intelligence (AI), a cross-party group of 40 MEPs in the European Parliament has called on the European Commission to strengthen the legislative proposals and include an outright ban on the use of facial recognition and other forms of biometric surveillance in public places.
The wide-ranging proposals, which were leaked ahead of their official publication, also outline tough new rules for ‘high-risk AI’. The latter includes algorithms used by the police service.
The proposed list of banned AI systems includes those designed or used in a manner that manipulates human behaviour, opinions or decisions (in turn causing a person to behave, form an opinion or take a decision to their detriment), those used for “indiscriminate surveillance” applied in a generalised manner, those employed for social scoring and those that exploit information or predictions and a person (or group of persons) to target their vulnerabilities.
Tony Porter QPM LLB, former Surveillance Camera Commissioner for England and Wales and now chief privacy officer at Corsight AI, has commented directly on the letter written by the MEPs for the Commission. He stated: “This signed letter and outright ban called for by a handful of MEPs fails to recognise that ‘biometric surveillance’, as it’s described, relies wholly on a ‘human in the loop’ strategy. Lawfully used, a facial recognition system is dependent upon human intervention. The software provides an indication or threshold as to the likelihood of a match, and the policies and processes accompanying that are aimed at ensuring an operator can conduct the surveillance lawfully.”
Porter continued: “It must be pointed out that, with most technology, facial recognition is rapidly advancing. Performance issues relating to speed and accuracy have massively improved over the last year. The ability to ‘train’ the software to eliminate bias is improving exponentially and the responsibility of developers to support operators in understanding the system is a clear guiding principle as to its effective and ethical use.”
Further, Porter observed: “The letter also fails to project a balanced argument as to the benefits and disadvantages. There’s no reference of the massive resource implications that can be saved, the dividend accrued by the public in the expeditious arrest and detention of the most serious offenders or the protection of the most vulnerable. Those commentators calling for an outright ban also have an ethical obligation to understand the technology, understand the level of public support it has gained and also understand the societal values it can deliver."
He added: “In an ever-increasingly complex and global world, crime and terrorism operate across borders, both nationally and internationally. Instances of child trafficking, sexual exploitation and drug trafficking are common between nations and this technology provides a tool in the armoury of our protective agencies. How can we possibly move towards the outright ban of a solution that can be used as a force for good and, ultimately, save lives?”
Strength of public concern
The European Commission is expected to present its final regulation on ‘high risk’ AI on Wednesday 21 April. A copy of the draft was leaked via Politico. The leaked draft doesn’t include a ban on the use of facial recognition or similar biometric remote identification technologies in public places, despite acknowledging the strength of public concern arising over the issue.
Commenting directly on the proposed rules for ‘high risk’ applications of AI, Porter explained: “I welcome the proposed EU regulations and the commitment of the current Government in its 2019 manifesto to provide clearer guidance as to the use of AI. This will surely strengthen public confidence in the use of facial recognition technology and other AI systems.”
He went on to outline: “We will see an explosion in the use of facial recognition technology across society from unlocking laptops through to facilitating access to cars, buildings and borders. Therefore, we must have confidence that this technology can be harnessed for the public good while, at the same time, developing a framework which is followed and clearly understood. To this end, our commitment to providing world-class technology is a driving force. If these proposed regulations can provide greater clarity then I firmly believe that’s a good thing for the industry.”
However, Porter noted: “The notion that AI software might be used in a manner that ‘manipulates human behaviour, opinions or decisions’ is quite abhorrent. At Corsight, we develop our software with the human operator in mind. Massive effort is expended in seeking to eliminate bias and ensuring that the ‘human in the loop’ understands the system, can manage the system and is in total command of the many and varied privacy settings that might change from deployment to deployment. Issues such as data retention, minimisation and purpose limitation sit comfortably alongside lawfulness, fairness and transparency. So, if the initial proposed regulation seeks to bring the rest up to the standards of the best, then that must only be a good thing.”
In conclusion, Porter said: “I’m confident that, if the EU continues to work with the AI industry, the developers and data scientists, while in parallel listening to the public, we will indeed ensure that this technology is seen to be a demonstrable force for good.”