Dr Deepthi Ratnayake, Principal Lecturer in Computer Science at University of Hertfordshire, and L. Fox Thomas MBCS, Vice Chair of BCS ISSG explore AI and its impact on women’s safety and security.
Artificial Intelligence (AI) quietly automates almost all of our everyday activities at speed. It is predicted that human-level AI, also known as Artificial General Intelligence (AGI) could be developed by 2040–2050, but regulating and enforcement of good practice in AI is still in its early days. The USA depends on The Blueprint for an AI Bill of Rights, whilst the UK is still seeking views and consultation for its white paper, ‘AI regulation: a pro-innovation approach’. Europe aims to be a global standard-setter with the its report, Special Committee on Artificial Intelligence in a Digital Age (AIDA), which aims to establish an AI roadmap up to 2030. The EU Parliament adopted recommendations of the AI Act and is in the process of reaching an agreement by EU members on the final form that this will take by the end of 2023.
Weak and narrow AI
AIs inherently have no legal personality or human conscience. This raises many issues including the picking up of biases as a result of data scraping, data lake pollution, the use of discriminatory algorithms, vague boundaries, undefined accountability of defective AIs, the potential to misdefine people, caging or controlling, segregation and isolation of the vulnerable. Intelligent systems have enormous potential to help humans flourish, however, they will never be fully realised if AI is not regulated and responsible use encouraged.
Is AI still a ‘boy's club’?
A Cambridge study reveals that only 22% of AI professionals are female, while the Harvard Business Review discusses many incidences of AI adopting gender bias from humans. A report from Nature revealed that a study using digital biomarkers has demonstrated that sex inequalities and gender biases have been seeping into the algorithms of biomedicine and healthcare.
Be part of something bigger, join BCS, The Chartered Institute for IT.
Where there is an overwhelming male dominance, there is bias inherent in the data that has been used, and this is being coded into algorithms. Women are often disadvantaged or put at risk by these unchecked algorithms. The invisibility of women has been widely debated, as has the idea that we are living in a world that has larger been created from the male perspective. Claims that women’s CVs are being binned by automated AI algorithms is also extremely sobering. Article 22 in UK GDPR, aimed to address automated profiling, but this does not appear to be enforced. How can we fix the issue if women are excluded from the decision making?
Frameworks and registrations
Medical practitioners are registered and must follow codes of practice. Patients therefore are able to trust the practitioners with their lives and health. Professionals are personally accountable for their practice and must always be prepared to justify decisions and actions. In the IT world, a greater degree of professional responsibility and accountability could go a long way to solidify women’s roles within the Secure Software Development Life Cycle (SSDLC), leading to employers or anyone else procuring coding services to be held accountable for the code being produced.
Some of the valuable suggestions from the practitioners working to resolve AI-related issues are:
- Champion a new campaign to align code developers with the health sector role framework
- Anyone developing, writing, reviewing code, should be registered and must adhere to an ethical code
- Legally recognise the skills and ethical behaviours expected in the coding profession, support coding as a professional career and not ‘just IT’
- Support for coders who have to refuse an employer’s request to develop code when the project that they have been asked to code for, is deemed unlawful or malicious
Joining BCS is about being recognised as a competent, ethical and accountable professional who is using tech for good in today's society. But the world needs to do more for women’s safety and security in the AI world.