T3 Podcast | Horizontal Intersections in Industries: Artificial Intelligence - Dr. M. Ayyüce Kızrak

In this episode of the T3 Podcast, we welcome Dr. M. Ayyüce Kızrak. The podcast covers a broad spectrum, from the historical development of artificial intelligence to its current applications, the differences between machine learning and deep learning, potential risks, and the future of research in Türkiye.

Dr. M. Ayyüce Kızrak notes that the definition of artificial intelligence varies depending on perspective. Its physical manifestations can be traced back to Al-Jazari’s robotic works, while its algorithmic foundations date to the 7th century. Today, the term refers to systems based on cognitive abilities and learning from data since its formal emergence in 1956. Artificial intelligence has now evolved from a research topic into a tool that simplifies everyday life.

Machine learning involves extracting features from data based on an expert’s experience, while deep learning, a subfield of machine learning, reduces the need for expert intervention by enabling the system to learn directly from data through artificial neural networks. Kızrak emphasizes that these concepts are interwoven structures and cannot be strictly separated from one another.

Artificial intelligence is integrated into many aspects of daily life, including facial recognition and personalized shopping recommendations. Its widespread use is illustrated through concrete examples such as fraud detection in banking, processing image and audio data in healthcare, and optimization in manufacturing and energy sectors.

Natural language processing models such as GPT-3 are considered revolutionary due to their ability to generate human-like text. Trained on massive datasets containing billions of words, these models can produce news articles and academic papers. Kızrak highlights the high economic value of this technology and stresses the need to encourage research efforts in Turkish.

Due to their “black box” nature, artificial intelligence systems often cannot explain their decision-making processes, leading to unpredictable errors and risks. This is particularly concerning in critical domains such as healthcare and law, where such failures can cause significant harm. To mitigate risks of discrimination arising from biases embedded in training data, it is recommended to develop testing environments with human oversight.

In recent years, Türkiye has seen a significant increase in motivation for artificial intelligence research. National policies such as the National Artificial Intelligence Strategy, the establishment of dedicated university programs, and growing incentives indicate that future research will accelerate further. Kızrak expresses confidence in the energy of young people in this field and asserts that this opportunity will not be missed as long as collaboration continues.

The most important advice for those seeking a career in this field is the ability to work across disciplines. Artificial intelligence is not a standalone domain but a horizontal field intersecting many sectors, and therefore everyone—from lawyers to engineers—must possess a certain level of awareness.