Governance after model speed links Riyadh discussions of the software addiction trap and technical sovereignty

Technical discussions in Riyadh are witnessing a notable shift in AI priorities, as the focus is no longer limited to what algorithms can achieve in speed and efficiency, but has moved to examining how they are built and used, placing mental well-being, cultural diversity and human interaction at the heart of ethical accountability. During the activities of the LIP conference and UNESCO’s Forum on AI Ethics, the perspectives of technology leaders, policy and ethics experts converged to study the challenges that accompany rapid development, foremost among them the risks of prolonged use, emotional dependence and the lack of fair representation of non-Western languages and identities.

In this context, Hani Shihabi, chief executive officer and co-founder of the platform “Thawra.ai” which he established with his brother Saeed Shihabi, presented a model of smaller-scale platforms that differentiate themselves from large corporations. The platform focuses on mental-health and cultural-diversity standards, based on the principle that technology should serve humans, not the reverse. Rather than training massive models from scratch at high cost, the platform relies on existing open-source models together with an internal “constitution”-like mechanism to steer their responses in line with its principles. Shihabi explained that the design deliberately limits excessive interaction, so the model delivers direct value to the user and ends the conversation without trying to keep them glued to the screen or chasing a longer interaction time.

This preoccupation with the risks of behavioral attachment receives direct support from international bodies.Dr. Dafna Vainholtz, head of the Ethics of Biology, Science and Technology division at UNESCO, warned against programming designed to capture attention and trap users in a continuous interaction loop, stating that prolonged reliance on chat-bot robots isolates individuals from natural social-communication patterns. Vainholtz emphasized the danger of replacing human emotional support with intelligent systems, reminding that friendship and human feelings go beyond mere exchange of automated texts. She affirmed that ethical frameworks to protect human rights and prevent harm must apply to both startups and large entities alike, with a need to focus on development and testing mechanisms rather than solely on market-launch pace.

The psychological dimension is also inseparable from the issue of sovereignty and cultural representation. While Vainholtz stressed the need for AI space not to be monopolized by a single language or culture, Shihabi affirmed that the growing base of Arab users on local platforms reflects a clear demand for region-rooted technologies that reduce continual reliance on Western infrastructure. These insights gain added weight as Saudi Arabia moves to develop its own tech ecosystem and to boost Arabic language and cultural identity in emerging national initiatives such as the “Humeen” project.

This shift in priorities imposes a practical reality on technology managers and software developers across the Gulf and the wider region. Rather than replicating Western consumption models that maximize time spent inside an app, regional systems are moving toward adopting open-source models governed by strict local controls that curb behavioral manipulation and lower training costs. If you lead a technical team, the next benchmark for evaluating your products will not stop at the model’s athletic performance, but will extend to how well you protect user privacy, prevent psychological drain, and are able to build authentic Arabic interfaces that go beyond superficial translation.