Google, Gemini -supported ASK Photos (ask photos) visual search feature could not maintain the distribution process as planned. According to Jamie Aspinall, one of the product administrators of Google Photos, the feature showed serious performance problems before it appeared before users. For this reason, the feature was presented to a limited number of users, but it was stopped before the distribution of a large scale. Aspinall, the problem, especially at the speed of the system, quality level and user experience intensified, he said.
Ask Photos’s main purpose was to enable users to make visual search using natural language in their Google Photos Archives. The Gemini Artificial Intelligence Model, which is located in the heart of the system, had the capacity to understand the questions and select the right visuals. However, it was not easy for this artificial intelligence technology to meet the expectations in the real world. During the tests, it was seen that the system works delayed, the quality of response remained below the expectation and the user interface could not offer the desired fluency.
Google postponed the distribution to improve Ask Photos experience
Accordingly, the company admitted that the current version of the feature is not suitable for widespread use. Ask Photos is currently kept open to a very limited group of user and the overall distribution process is suspended. Google has announced that this decision is temporary and that the new, improved version will be presented within two weeks. The new version aims to increase both reaction time and content accuracy.
Ask Photos’s first presentation was made at the 2023 Google I/O event. Then, in September, he was introduced to some users within the scope of early access. Since then, warnings about the deficiencies as well as positive comments about the system had come up. Nevertheless, Google chose not to present this feature to the general use without making it more stable.
In addition, Google’s artificial intelligence -based services also draws attention to experience similar problems. The AI OverViews feature presented in the search results was criticized for faulty and misleading content. The visual production tool also brought about reliability discussions by producing unexpected results. It is frequently questioned whether the company has applied adequate quality control processes during the transition to artificial intelligence.
In addition to all these, the criticisms from the company and from the old employees also signals that the process does not work healthy. Although Google’s orientation of artificial intelligence has accelerated, this speed is thought to have negative effects on the user experience. The stopping of Ask Photos early from expected also looks like a current example of this painting. It is clear that more solid foundations based on solid foundations should be developed in order to gain the trust of users.
Although Google seems determined to quickly integrate artificial intelligence characteristics, technical difficulties suggest that this integration is hurried. In particular, considering the sensitivities of users’ photo archives, incorrect results may pose a risk for both safety and experience. Therefore, the expansion of the features in limited test groups can prepare the ground for the repetition of such problems. How Google will learn from this process is another issue.
What innovations of developed Ask Photos will be discussed. The interface is expected to design more simple and directing the user. On the other hand, it is aimed to better analyze natural language in the form of response of artificial intelligence. The quick, accurate and meaningful access to the visual that the user is looking for will be decisive in terms of the success of the new version.